I have pretty different opinions on this I guess.

I think QA/Testing is definitely fading away. It'll never be gone, but its not going to grow anything at all like general software development. And its going to be less and less compensated relative to development except for rare circumstances.

The move to SaaS (or even just things like web apps - anything that doesn't require sending code to clients to install themselves), better deployment practices (red/green, feature flags, canary systems, etc) and better "observability" tools (infrastructure and code) are all working against the traditional QA role.

The cost of finding a bug in Production is a lot lower than it use to be (don't need bug reports or user submitted stack traces, can be found before affecting the majority of users, etc.) and the time to fix the bug is way lower (if you're doing multiple releases/day then a targeted fix is just one more release among many).

Other trends like developers actually owning, releasing, and supporting their code in production also work against traditional QA roles since developers have much higher incentives to release good/stable code.

There will always be companies/applications/industries where the cost of having a failure (or fixing a failure) is really high - and it makes a lot of sense in those cases to have a strong test/QA team. But I don't think this is going to be a growing proportionately to the rest of the software industry.

There may also be roles/spots for sort of "internal support engineers" that build tooling/automation around things like user creation / test data / etc. But I think the value here will be a lot lower than in historical QA roles where knowing requirements/edge cases/etc was super important. It's basically useful where you can find less skilled people to do work that makes your more skilled people more efficient.


This seems like a really good reason NOT to have testers. Developers not understanding requirements is not a recipe for good code (or a good product), imo.

Like any job title, I'm sure this has ended up with a wide range of meanings/responsibilities, but the fashionable SRE role is very different from QA. It's like a dev/ops mix and has a pretty strong future (imo).

IDK. My current role doesn't have a QA step, which surprised me very much when I joined in January to lead the rewrite of v1. TBH I don't miss it a ton, especially since, to be blunt, I haven't had a lot of success with offshore QA and this would have been offshore for sure. However we have the standard "how do we know things are done" issues that you'll get without good QA and we accept things that for sure need rework because of it, wasting everyone's time.

Interestingly the history of our product is that the v1 rollout was a nightmare and for sure if they had real QA it would have been better. Hard to say.

Our situation was that there was no living documentation of everything the site did or all the flows. When we finally tried to document just login flow (basically they put a new business analyst on it), it filled up a 20' wall with printouts of 8"x11" screenshots.

Which sure is a problem. But maintaining those kind of living docs through endless small and large change requests and bugfixes would have also been a gigantic job that apparently no one wanted to do.

So in practice you might get a new dev on a project who understands the scope of work they're working on, but not the big picture and all the use cases that the work might touch.

As an example: we might have some old dwindling class of legacy user with some weird product that we don't even make anymore. Nobody really thinks about them much but you can sure as hell bet if you break their stuff you're gonna get calls.

Now - maybe every dev should spend a lot of time setting up these legacy users and testing. But 99% of the time your stuff isn't going to break their flow. So you spent all that time and effort for nothing. Whereas QA already has these users sitting around and it's much easier for them to test against them. Multiply this by dozens of weird edge cases. Obviously a massively good case for automation.

Basically you wind up in a lot of sticky situations when you develop a monster website over a decade pretty fast and loose, which you can do since your bugs don't generally kill people or cost tons of money. And again - also being the tail on a giant dog of legacy back end systems that you depend on but have no control over.

My new job has no QA. I think my most recent one is the only one where devs didn't do their own testing. It's just a whole different way of developing. I'm trying to think of a good analogy. It's something like being able to fire off a quick sports column and have someone else fix all your spelling and grammatical errors.

I also found out yesterday that for a big chunk of our products, devs are testing some then deploying straight to production. Note: this app takes donations from alumni. So if it breaks they can lose a lot of money. That would terrify me as a dev.

But the ironic thing is we're already halfway to DevOps nirvana. We've eliminated the DEV handing off to OPS step. Now we just need to test and automate everything. What's old is new, everything comes around, etc.

LOLOLOL this AWS Summit is kind of a joke. First speaker went through 60 slides in 14 minutes (out of allotted 30) and was pretty much just her rambling thoughts on "what is lambda?".

Second speaker can't get the AV working. JJ can you please get down here and explain to these people how this **** isn't that hard and they need to step up their game?

Yea kinda disappointing.

In their defence, organizing AV for a conference is not at all the same as organizing a video conference and has a lot more challenges.

Although honestly I’m surprised because usually AWS conferences have this **** down with a bunch of rehearsals and stuff. Summits might be different though.

It’s like a big sales pitch

The keynotes/main presentations at conferences usually suck, imo. Sometimes you'll get good announcements / cool stuff - but that's kind of rare and definitely rare at like regional conferences like AWS Summits.

Any good actual talks/presentations?

I thought the keynote was fine. It’s cool to see use cases of a lot of the products I know nothing about.

The use cases were great, the main speaker i didn’t like

Suzzer, this alexa for business meeting is about making video conferencing easier. Lol

Lolololil

Jmakin - if you’ve been, how would you compare this crowd to WSOP?

More indian

and speak of the devil.

Went to a meeting today. Speaker couldn't get skype to work on his mac and him and two other guys took 20 minutes to get going. Around 40-50 people in the meeting. A man day of production down the toilet.

I would estimate "time lost to video conferencing issues" at my company at 0. I guess our IT department knows their **** more than most?

We use Zoom + Macbooks. Every conference room has a camera and TV and Zoom's manager app running on a dedicated iPad. I don't schedule a lot of meetings myself but for those who do, it's apparently very easy to get a Zoom meeting attached to a gcal event because almost all of my meetings have a Zoom link in the calendar, attached to the right conference room. Someone physically in the room presses "start" on the Zoom (which already has the meeting schedule for that room) and away we go. Magic!

Witchcraft!

This suzzer/jmakin liveblog was some top tier stuff!

We need more people to go to some conferences together and post about it itt.

Sounds like Zoom + dedicated iPad is far superior to the massively expensive dedicated hardware that big companies were purchasing in the last few decades. Not surprising.

In a big company you usually have a mix of mac and PC of various stripes. Could complicate things.

We use Lync 2011

Dude is literally walking around like this:


(not actual pic)

I'm like top 10% attractiveness at this conference - which is extremely hard to do. It's like driving through backwoods Arkansas or something - except smart ugly people instead of dumb.

Have I ticked off enough offensive boxes yet?

Yea this is a sausage fest

I can confirm that jmakin is a handsome gentleman - like top 1% here.

The few women here aren't too bad. Some hotties.

That is not saying a lot for the crowd.

I left at 3:30 because none of it was really applicable to what we do. My coworkers were happy to get a day off but all I could think about was losing a day of productivity from 3 people. Lol